Memory including error correction circuit and operating method thereof

Abstract

Disclosed is an operating method of a memory, and the operating method may include reading, from selected memory cells included in the memory, codewords including data and an error correction code; detecting errors in the codewords; correcting the errors in the codewords; re-writing the error-corrected codewords to the selected memory cells; re-reading the re-written error-corrected codewords from the selected memory cells; and determining whether the errors are permanent errors in response to a determination that an error is present in the re-read error-corrected codewords.

What is claimed is: 
     
       1. An operating method of a memory, the operating method comprising:
 reading, from selected memory cells included in the memory, codewords including data and an error correction code; 
 detecting errors in the codewords; 
 correcting the errors in the codewords; 
 re-writing the error-corrected codewords to the selected memory cells; 
 re-reading the re-written error-corrected codewords from the selected memory cells; 
 determining whether the errors are permanent errors in response to a determination that an error is present in the re-read error-corrected codewords; 
 classifying a row where a greatest number of errors are detected among rows of the memory as a first candidate of a bad region, while preferentially classifying a row including the permanent errors among rows having a same greatest number of errors as the first candidate of the bad region as a second candidate of the bad region; and 
 classifying the second candidate of the bad region as the bad region when the same greatest number of errors of the second candidate of the bad region is greater than or equal to a threshold value. 
 
     
     
       2. The operating method of  claim 1 , wherein the determining of whether the errors are permanent errors includes:
 determining that the errors are the permanent errors when an error is present in the re-read error-corrected codewords, and 
 determining that the errors are transient errors when an error is not present in the re-read error-corrected codewords. 
 
     
     
       3. The operating method of  claim 1 ,
 wherein the reading of the codewords, the detecting of the errors, the correcting of the errors, the re-writing of the error-corrected codewords, the re-reading of the re-written error-corrected codewords and the determining of whether the errors are permanent errors are performed during an error check operation of the memory, and 
 wherein the bad region is classified in the error check operation. 
 
     
     
       4. The operating method of  claim 1 , further comprising storing, when the bad region is classified, a row address of the bad region and whether the permanent errors are included in the bad region. 
     
     
       5. An operating method of a memory, the operating method comprising:
 generating an internal row address; 
 generating an internal column address; 
 reading, from memory cells selected according to the internal row address and the internal column address, codewords including data and an error correction code; 
 detecting errors in the codewords; 
 increasing an error count for a region corresponding to the internal row address, in response to the detection of the errors; 
 correcting the errors in the codewords; 
 re-writing the error-corrected codewords to the selected memory cells; 
 re-reading the re-written error-corrected codewords from the selected memory cells; and 
 determining whether the errors are permanent errors in response to a determination that an error is present in the re-read error-corrected codewords, and activating a permanent flag representing that the errors are the permanent errors; 
 classifying a row with a higher error count as a first candidate of a bad region, while preferentially classifying a row for which the permanent flag is activated among rows having a same greatest number of errors as the first candidate of the bad region as a second candidate of the bad region; and 
 classifying the second candidate of the bad region as the bad region when the greatest number of errors of the second candidate of the bad region is equal to or greater than a threshold value. 
 
     
     
       6. The operating method of  claim 5 ,
 wherein the generating of the internal row address, the generating of the internal column address, the reading of the codewords, the detecting of the errors, the increasing of the error count, the correcting of the errors, the re-writing of the error-corrected codewords, the determining of whether the errors are permanent errors are performed during an error check operation of the memory, and 
 wherein the error count and the permanent flag are assigned to each row of the memory during the error check operation. 
 
     
     
       7. The operating method of  claim 5 , further comprising registering, when the bad region is classified, a row address of the bad region and whether the permanent errors are included in the bad region. 
     
     
       8. The operating method of  claim 7 , further comprising registering, when the bad region is classified, whether the permanent errors are detected twice or more in a same region. 
     
     
       9. A memory comprising:
 a memory core; 
 an error correction circuit configured to detect and correct errors in codewords including an error correction code and data, read from the memory core; 
 an error check address generation circuit configured to generate a row address and a column address to be used for an error check operation; and 
 an error check operation control circuit configured to: 
 control the error check operation for classifying a bad region of the memory core, 
 re-write, in response to the detection of the errors in the codewords read from memory cells selected by the row address and the column address in the memory core, the error-corrected codewords to the selected memory cells, 
 re-read the error-corrected codewords from the selected memory cells, and 
 determine whether the errors are permanent errors in response to whether an error is present in the re-read error-corrected codewords; and 
 an error log circuit configured to classify a row where a greatest number of errors are detected among rows of the memory as a first candidate of a bad region, while preferentially classifying a row including the permanent errors among rows having a same greatest number of errors as the first candidate of the bad region as a second candidate of the bad region, and classify the second candidate of the bad region as the bad region when the same greatest number of errors of the second candidate of the bad region is greater than or equal to a threshold value. 
 
     
     
       10. The memory of  claim 9 , wherein the error check operation control circuit:
 determines that the errors are the permanent errors when an error is present in the re-read error-corrected codewords, and 
 determines that the errors are transient errors when an error is not present in the re-read error-corrected codewords. 
 
     
     
       11. The memory of  claim 9 , further comprising:
 an error count circuit configured to record, for each row of the memory core, a number of detected errors and a permanent flag representing whether the errors are the permanent errors; and 
 the error log circuit configured to classify and register the bad region by using information recorded in the error count circuit. 
 
     
     
       12. The memory of  claim 9 , wherein the error log circuit stores a row address of the bad region and the number of detected errors for the bad region. 
     
     
       13. The memory of  claim 12 , wherein the error log circuit further stores information representing whether the permanent errors are present in the bad region.